Re: Not Much LINUX Stuff of Late

On 2026-07-31 13:37, Nuno Silva wrote:
> On 2026-07-29, Carlos E. R. wrote:
> 
>> On 2026-07-29 17:04, John Ames wrote:
>>> On Wed, 29 Jul 2026 08:25:18 -0400
>>> Chris Ahlstrom <[email protected]> wrote:
>>>
>>>> And systemd seems to be to be a good thing. It sure minimizes boot
>>>> times.
>>>
>>> Does anybody really care about boot times anymore? *nix uptimes can
>>> generally be measured in months or years,
>>
>> Huh, no. Only servers do that. My laptop, for instance, would be
>> hibernated or suspended twice per day; and as to uptime, it crashes
>> during wake up some times.
> 
> As soon as you hibernate, I'd think the benefits reportedly gained by
> e.g. systemd are less relevant, or not relevant at all, as the system
> wouldn't go through the boot process, but would rather be restored in an
> already booted state.

On restore from hibernation, the kernel boots till the moment it can 
read the swap and see an hibernation image there. Then it loads it 
appropriately. At some point, systemd intervenes and does things. Before 
systemd, there were scripts that also did things. Many scripts, 
actually. I don't remember the name of the package. The obvious is 
restarting the network or checking what is connected externally.

> 
> As for improvements, I'm quite happy with the speed and sequentiality I
> can achieve with OpenRC, but of course there's no reason why there can't
> be faster approaches and parallelization for cases where it's
> needed/desired, I think OpenRC also does some parallelization too if
> desired, for example.
> 
>>> and off an SSD even my pokey
>>> little Eee 904 (running Devuan with sysvinit) goes from zero to display
>>> manager in less than a minute.
> 
> I can't forget the moment a couple decades ago where I realized my
> Pentium II desktop was booting into the login prompt much, much faster
> than a Windows Vista laptop took to get to its login screen... and that
> was a Pentium II with less performant I/O than more current hardware. It
> might be obvious in hindsight, perhaps especially given the absence of a
> graphical login manager, but still fascinating to realize.
> 

I was happy with initd at the time. Now it is systemd, and I am also 
happy. I don't really care about the boot speed: my desktop machine 
takes minutes to boot. Before, if I wanted to create a new service, I 
looked for a similar script and adapted it for the new. Now I do the 
same with a service file.

The maintainers are happy, they have their reasons, that's good enough 
for me.

-- 
Cheers,
        Carlos E.R.
        ES🇪🇸, EU🇪🇺;